RFR: JDK-8305962: update jcstress to 0.17

[Aleksey Shipilev]

On Fri, 30 Jun 2023 20:09:50 GMT, Leonid Mesnik <lmesnik at openjdk.org> wrote:

> The fix changes jcstress version and update some parameters used by the jtreg wrapper.

Not actually sure if modern jcstress can be executed with this wrapper, given how it might need to drop in JNA blobs before execution.

There is a (not yet released) version of jcstress that gives `-tb` (time budget) option that might be useful here.

test/hotspot/jtreg/applications/jcstress/JcstressRunner.java line 45:

> 43:  */
> 44: @Artifact(organization = "org.openjdk.jcstress", name = "jcstress-tests-all",
> 45:         revision = "0.16", extension = "jar", unpack = false)

This is is not `0.17`, as issue title implies.

test/hotspot/jtreg/applications/jcstress/JcstressRunner.java line 62:

> 60:                             + JcstressRunner.class.getName(), e);
> 61:         }
> 62:         return artifacts.get("org.openjdk.jcstress.jcstress-tests-all-0.16")

Not `0.17` either.

test/hotspot/jtreg/applications/jcstress/JcstressRunner.java line 114:

> 112:         // The "default" preset might take days for some tests
> 113:         // so use sanity testing by default.
> 114:         String mode = "sanity";

`sanity` mode is incorrect for actual testing runs. Should be at least `quick`.

test/hotspot/jtreg/applications/jcstress/JcstressRunner.java line 129:

> 127: 
> 128:         extraFlags.add("-sc");
> 129:         extraFlags.add("false");

Is this to shorten the execution time? I'd recommend `-af GLOBAL` too then, since we are reducing the test matrix for it anyway.

-------------

Changes requested by shade (Reviewer).

PR Review: https://git.openjdk.org/jdk/pull/14742#pullrequestreview-1514173614
PR Review Comment: https://git.openjdk.org/jdk/pull/14742#discussion_r1252885192
PR Review Comment: https://git.openjdk.org/jdk/pull/14742#discussion_r1252885350
PR Review Comment: https://git.openjdk.org/jdk/pull/14742#discussion_r1252885661
PR Review Comment: https://git.openjdk.org/jdk/pull/14742#discussion_r1252890456